Original Research

Comparison Between Back Squat, Romanian

Deadlift, and Barbell Hip Thrust for Leg and Hip
Muscle Activities During Hip Extension
Jose Delgado,1 Eric J. Drinkwater,1,2 Harry G. Banyard,1,3,4 G. Gregory Haff,1,5 and Kazunori Nosaka1
1
Center for Exercise and Sports Science Research, School of Medical and Health Sciences, Edith Cowan University, Joondalup,
Western Australia, Australia; 2Center for Sport Research, School of Exercise & Nutrition Sciences, Deakin University, Melbourne,
Victoria, Australia; 3West Australian Football Commission, Subiaco, Western Australia, Australia; 4Department of Health and Medical
Sciences, Swinburne University of Technology, Melbourne, Victoria, Australia; and 5Directorate of Sport, Exercise and Physiotherapy,
University of Salford, Greater Manchester, UK

Abstract
Delgado, J, Drinkwater, EJ, Banyard, HG, Haff, GG, and Nosaka, K. Comparison between back squat, Romanian deadlift, and
barbell hip thrust for leg and hip muscle activities during hip extension. J Strength Cond Res 33(10): 2595–2601, 2019—This study
compared muscle activities of vastus lateralis (VL), biceps femoris (BF), and gluteus maximus (GM) during the back squat (SQ),
Romanian deadlift (RDL), and barbell hip thrust (BHT) exercises performed with the same load (60 kg) and at one repetition
maximum (1RM). Eight men with a minimum of 1 year’s lower-body strength training experience performed the exercises in
randomized order. Before each exercise, surface electromyography (EMG) was recorded during a maximal voluntary isometric
contraction (MVIC) and then used to normalize to each muscle’s EMG during each trial. Barbell hip thrust showed higher GM activity
than the SQ (effect size [ES] 5 1.39, p 5 0.038) but was not significantly different from RDL (ES 5 0.49, p 5 0.285) at 1RM. Vastus
lateralis activity at 1RM during the SQ was significantly greater than RDL (ES 5 1.36, p 5 0.002) and BHT (ES 5 2.27, p 5 0.009).
Gluteus maximus activity was higher during MVIC when compared with the 60 kg load for the SQ (ES 5 1.29, p 5 0.002) and RDL
(ES 5 1.16, p 5 0.006) but was similar for the BHT (ES 5 0.22, p 5 0.523). There were no significant differences in GM (ES 5 0.35,
p 5 0.215) and BF activities (ES 5 0.16, p 5 0.791) between 1RM and MVIC for the SQ. These findings show that the RDL was
equally as effective as the BHT for isolating the hip extensors, while the SQ simultaneously activated the hip and knee extensors.

Key Words: gluteus maximus, biceps femoris, vastus lateralis, EMG, root mean square

Introduction Previous research by Bishop et al. (3) examined the effects of an

8-week BHT strength program on sprint performance and reported
Hip extension force is important for accelerating the body’s up-
that the use of the BHT did not transfer into improvements in sprint
ward and forward movement such as during sprinting and
performance. Conversely, Contreras et al. (9) favored the BHT for
jumping (18). To increase force production during hip extension,
improving 10- and 20-m sprint times over the front squat, although
the back squat (SQ), Romanian deadlift (RDL), and barbell hip
it is important to note that the subjects in the study had no previous
thrust (BHT) are often used in training. The SQ involves moving
experience with the BHT but had 1 year of SQ experience.
both the hip and knee through flexion to full extension, eliciting
Therefore, superiority of training adaptation for the BHT over the
substantial activation of the hip and knee musculature (5). The
front squat would be expected, as the introduction of a novel
RDL is commonly used with the aim of strengthening the gluteal
training stimulus (i.e., BHT) is likely to elicit higher adaptation over
and hamstring muscle groups as well as the spinal erectors
6 weeks than continuing a stimulus to which they were already
(19,20). The RDL has also been considered to be crucial for the
accustomed (i.e., SQ). In addition, Andersen et al. (2) showed
development of weightlifting movements such as the clean and
higher GM activation during the BHT when compared with the
snatch (12). It has been suggested by Contreras et al. (7) that the
barbell deadlift and hex bar deadlift, with superior biceps femoris
BHT is superior to the SQ in eliciting higher gluteal muscle ac-
(BF) activity seen during the barbell deadlift. Further study by
tivity, developing terminal hip extension strength in the gluteus
Contreras et al. (8) also reported that the peak activity of the upper
maximus (GM) thereby increasing horizontal force production,
and lower GM and BF was greater during the BHT when compared
and increasing the contribution of the GM relative to the ham-
with the SQ, but no significant differences were found between the
strings during hip extension movement. Although the BHT has
exercises for peak vastus lateralis (VL) activity.
gained recent popularity, there is a lack of evidence to support the
While the study by Contreras et al. (8) reported differences in
notion that the BHT does in fact increase horizontal force pro-
muscle activation between the SQ and BHT, there seems to be
duction of the hip or elicits higher GM activity than the SQ or
several limitations to the methodology used. First, Contreras et al.
RDL in athletic populations.
(8) recorded electromyography (EMG) from the upper and lower
Address correspondence to Jose Delgado, joseddm@hotmail.com. GM. However, these sites do not conform to the Surface EMG for
Journal of Strength and Conditioning Research 33(10)/2595–2601 Non-Invasive Assessment of Muscles (SENIAM) guidelines for
ª 2019 National Strength and Conditioning Association collecting EMG data (13). Second, “manual resistance” was used

2595

Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Back Squat, Romanian Deadlift and Barbell Hip Thrust (2019) 33:10
during the normalization protocol for the VL, as well as
a “standing glute squeeze” to normalize GM EMG signals, but it
was unlikely that an adequate fixation necessary for maximal
activation of the muscles involved was achieved (14). Further-
more, Contreras et al. (8) had subjects performing 10 repetitions
with a predicted 10 repetition maximum (10RM), which is esti-
mated to be approximately 75% of the 1RM (10), thus it seems
unlikely that maximal muscle activation was achieved during any
exercise. It is necessary to compare the SQ and BHT during an
actual 1RM using SENIAM guidelines, providing an indication of
muscular activities during maximal effort movement. Moreover,
as the BHT provides a movement with limited range, there are
usually higher loads associated with its 1RM and would therefore
have a relatively higher load for the same relative intensity, such
as 60% of 1RM when compared with the SQ and RDL. In-
vestigating muscle activation differences with the same absolute
standardized submaximal load (e.g., 60 kg) across exercises
would highlight activation differences without load variation. It is
also important to assess muscle activities during a standardized
maximal voluntary isometric contraction (MVIC) to normalize
the muscle activities recorded during exercise (17).
Therefore, the aim of this study was to compare the SQ, RDL,
and BHT at 1RM and at a standardized load (60 kg) for muscle
activities of VL, BF, and GM during hip extension. It was hy-
pothesized that the SQ would elicit similar GM and BF activation
when compared with the RDL and BHT at 1RM and 60 kg, while
VL activity would be higher in the SQ when compared with those
of the RDL and BHT at 1RM and 60 kg.
Methods
Experimental Approach to the Problem
Electromyography of the hip and knee extensors was recorded
during the SQ, BHT, and RDL while subjects lifted a standardized
60 kg and exercise-specific 1RM load. Subjects were asked to
attend one familiarization session followed 3–7 days later by a 3-
hour testing session to compare the SQ, RDL, and BHT for EMG
activities of VL, BF, and GM.
Electrodes were placed on the GM, BF, and VL according to the
SENIAM guidelines (13). After this, subjects performed a 5-
minute standardized warm-up on a stationary cycle ergometer
and 20 body mass squats. This was followed by an EMG nor-
malization protocol on an isokinetic dynamometer (Biodex Sys-
tem 3 Pro; Biodex Medical System, Shirley, NY, USA) to record
EMG during a maximal voluntary isometric contraction (MVIC).
After the normalization session, subjects were given a 10-minute
rest before performing the exercises (i.e., BHT, SQ, and RDL) to
1RM, including a warm-up set with a submaximal load of 60 kg
and a 10-minute rest between exercises. Electromyography from
the 3 muscles was continuously recorded during these trials.
Subjects
Eight resistance-trained men who had a minimum of 1 year of
lower-body resistance-training experience, and could perform an
SQ with a load of at least 150% of their body mass to a depth in
which the top surface of the thighs was lower than the top of the
knees, were recruited. In the familiarization session, the subjects
performed 10 repetitions of each exercise with a standard 20 kg
weightlifting barbell to ensure proper lifting technique of each
exercise. The sample size of 8 subjects was estimated using
G*Power 3.1 (11) for a within-subject repeated-measures analysis
2596
Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
of variance (ANOVA) (a 5 0.05, b 5 0.80) to detect differences
(effect size [ES] 5 0.40) in EMG activity of GM, VL, and BF for
the 3 different exercises (i.e., BHT, SQ, and RDL) over 2 in-
tensities (60 kg and 1RM). The average 6 SD (range) age, height,
and body mass of the subjects were 25.0 6 3.3 (18–30) years,
177.9 6 6.6 (169–177) cm, and 84.0 6 6.5 (74.5–91.5) kg, re-
spectively. Subjects abstained from their regular training between
3 and 7 days before the testing sessions. The experimental pro-
cedures were approved by Edith Cowan University’s Human
Research Ethics Committee and were in agreement with the
principles of the Declaration of Helsinki. The subjects were
briefed on the study procedures and informed consent was
obtained in writing before participating in the study and were
informed that they could withdraw from the study at any time.
Procedures
One Repetition Maximum Assessment. Every testing session was
supervised by a Certified Strength and Conditioning Specialist to
ensure proper lifting technique under maximal loads. Each subject
performed the 3 exercises inside a power rack with safety holes
spaced 2.54 cm apart, by following the National Strength and
Conditioning Association protocol (10). A warm-up with light
resistance that easily allowed 10 repetitions was performed. After
a 1-minute rest period, a load of 15–20 kg was added to the barbell
and each subject was instructed to perform 3–5 repetitions, fol-
lowed by a 2- to 4-minute rest period. Load was gradually in-
creased, and this process was repeated until a body mass was
reached where failure of technique occurred, or the subject was not
able to perform more than one repetition. There were 3–4 attempts
to identify maximal weight lifted. The 1RM was recorded as the
highest successfully lifted barbell load with correct technique.
For the SQ, each subject was asked to place the bar over the
upper trapezius in a high-bar position and continue to flex the
hips and knees until the top of the thighs were lower than the top
of the knees, maintaining a neutral spine alignment. For the BHT,
peak barbell height was recorded at the point of maximal hip
extension during the initial repetitions of the warm-up by placing
a wooden stick through the highest hole in which the safety pins
insert into the power rack. Each subject was asked to thrust the
barbell to this height during each test and received constant
feedback to ensure there was no barbell tilt at the top. This height
was kept constant by requiring the subject to hit the wooden stick
with the barbell every repetition. One repetition maximum was
recorded as the highest lifted load until he could not reach peak
barbell height at the end of the concentric action. Each subject
was asked to place his upper back on a bench 40-cm high, with the
barbell placed at the crease of his hips with a barbell pad to
minimize discomfort. The subjects were instructed to extend
through the hips, with their feet firmly planted on the ground. For
the RDL, the bar was placed on a rack set at the subject’s hip level.
Each subject was asked to use weightlifting straps to prevent his
grip from being a limiting factor. They were instructed to stand
with the weight, take 2 steps back out of the rack, and perform an
RDL while maintaining a neutral spinal alignment and extending
through the hip. Range of motion in the RDL was standardized to
the barbell reaching the bottom of the patella for every subject.
One repetition maximum was recorded as the highest lifted load
with correct technique or until the subject could not reach peak
barbell height at the end of the concentric (upward) motion.
Electromyography Recording and Analyses. All skin preparation
procedures were followed according to the SENIAM guidelines,
 Back Squat, Romanian Deadlift and Barbell Hip Thrust (2019) 33:10 | www.nsca.com

and each electrode was tested to ensure a reading of less than 5 kV. kg, and 164.1 6 13.9 (135–180) kg, respectively (Table 1). The
Electromyography signals were recorded using pairs of silver chlo- subjects lifted lighter loads for the SQ when compared with the
ride surface electrodes (2-cm diameter, Noraxon Dual Electrodes; RDL (22.9 6 17.8 kg, ES 5 1.19. p 5 0.019) and the BHT (20.1
Noraxon USA, Inc., Noraxon, Scottsdale, Arizona, US) through 6 15.4 kg; ES 5 1.05. p 5 0.018). The 60 kg load relative to the
a wireless EMG system (Zero Wire System; Aurion, Milano, Italy) 1RM load was 42.3 6 5.5% for the SQ, 36.4 6 4.2% for the
recording at 2,000 Hz using a telemetry transmitter (Wave Wireless; RDL, and 36.8 6 3.5% for the BHT.
Cometa Systems, Milan, Italy), which was analyzed using LabChart
8 software (PowerLab system, version 6.1.3; ADInstruments, NSW,
Australia). All electrodes were placed on each subject’s dominant Gluteus Maximus Electromyography
side. For the VL, electrodes were placed at two-thirds of a line As shown in Figure 1, the RMS of the SQ for the 60 kg load was
measured by a measuring tape, connecting the anterior superior iliac smaller than that of the BHT (ES 5 1.36, p 5 0.008), but not
spine to the lateral side of the patella, in the direction of the muscle significantly different from that of the RDL (ES 5 0.37, p 5
fibers. Electrodes were placed at 50% of a line measured by a mea- 0.325). For the 1RM load, no significant differences in RMS
suring tape, connecting the ischial tuberosity and the lateral epi- were evident between the RDL and SQ (ES 5 0.57, p 5 0.127)
condyle of the tibia for the BF, and at 50% of a line connecting the as well as the RDL and BHT (ES 5 0.49, p 5 0.285), but the
sacral vertebrae and the greater trochanter for the GM. RMS of BHT was greater than that of SQ (ES 5 1.39, p 5
To normalize EMG activity of the VL, BF, and GM during the 0.038). When comparing with the RMS during MVIC, the SQ
SQ, RDL, and BHT, EMG activity of each muscle during a maximal (ES 5 1.29, p 5 0.002) and RDL (ES 5 1.16, p 5 0.006)
voluntary isometric contraction (MVIC) was recorded based on showed smaller RMS during the 60 kg load, but the BHT
Konrad’s (14) proposed test positions and comprised 3 MVICs for showed no such difference (ES 5 0.22, p 5 0.523). There were
each muscle group individually, to assess maximal EMG activity. no significant differences in RMS when comparing the MVIC
Electromyography root mean square (RMS) during an MVIC was and the 1RM for the SQ, RDL, and BHT (ES 5 0.35, p 5
recorded during a one second window at peak torque produced in 0.215; ES 5 0.08, p 5 0.806; and ES 5 0.30, p 5 0.375,
the MVIC. Electromyography RMS was calculated using LabChart respectively).
8 software and was recorded during the concentric phase of each
exercise at 1RM and 60 kg. Differences in maximal activation
(MVIC to 1RM) and changes between the 3 exercises were de- Biceps Femoris Electromyography
termined during the concentric phase of each exercise. Each subject No significant differences in the RMS of the BF were observed
was seated and strapped to the isokinetic dynamometer with a hip between the SQ and BHT (ES 5 0.98, p 5 0.075) as well as the SQ
angle of 90° and was asked to maximally extend and flex for 3 and RDL (ES 5 0.35, p 5 0.326) for the 60 kg load (Figure 2). In
repetitions at 3 different positions. For the VL, subjects were asked addition, no significant differences were evident when comparing
to maximally extend their knee fixed at 70° of flexion (where 0° is the BHT with the RDL (ES 5 0.71, p 5 0.078) for the 60 kg load.
the knee fully extended). For the BF, the subjects were asked to For the 1RM load, no differences were observed when comparing
maximally flex their knee set at the same angle of 70° of flexion. the SQ to BHT and RDL (ES 5 0.69, p 5 0.239 and ES 5 0.38, p
Finally, for the GM, subjects were asked to lie prone and maximally 5 0.181, respectively), and BHT to RDL (ES 5 0.33, p 5 0.508).
extend their hip with their upper leg fixed at 180° of hip flexion When comparing with MVIC, no differences in RMS were found
(i.e., full hip extension). They were given 1-minute rest between for the SQ (60 kg: ES 5 0.69, p 5 0.101, 1RM: ES 5 0.16, p 5
MVICs and 2–3 minutes of rest between muscle groups. 0.791), RDL (60 kg: ES 5 0.54, p 5 0.124, 1RM: ES 5 0.60, p 5
0.308), and the BHT at 60 kg (ES 5 0.14, p 5 0.608), but higher
BF activity was observed during the BHT at 1RM (ES 5 0.99, p 5
Statistical Analyses 0.001).
Data analysis was completed in SPSS statistical analysis soft-
ware (Version 25; IBM Corporation, Armonk, NY, USA). One-
way ANOVAs with repeated measures were used to analyze for
differences in EMG RMS for each muscle group (i.e., GM, VL,
and BF) between the 3 different exercises with maximal loads
(1RM) as well as the intensities (%1RM) lifted during a 60 kg Table 1
repetition. Where a statistically significant effect was identified, One repetition maximum load (1RM) and its relative value to the
univariate statistics (least significant differences) were used to body mass (1RM/BM) of back squat, Romanian deadlift, and
barbell hip thrust for each subject (A–H), and the mean and SD of 8
identify between which pairs statistically significant differences
subjects.
lay. Differences were determined to be statistically significant at
p # 0.05. Results are expressed as mean difference 695% Back squat Romanian deadlift Barbell hip thrust
confidence interval, Hedge’s g ES, and p value. Qualitative Subject 1RM (kg) 1RM/BM 1RM (kg) 1RM/BM 1RM (kg) 1RM/BM
descriptors of standardized ES were assessed using the criteria: A 130 1.44 150 1.64 170 1.86
trivial, 0.0–0.19; small, 0.2–0.59; moderate, 0.6–0.79; and B 180 2.00 170 1.87 180 1.98
large, .0.80 (6). C 145 1.81 170 2.13 153 1.91
D 155 1.72 155 1.72 165 1.83
E 145 1.81 180 2.24 170 2.11
Results F 130 1.67 155 1.97 135 1.71
G 117 1.57 145 1.95 170 2.28
Exercise H 150 1.75 210 2.45 170 1.98
The average 6 SD (range) 1RM load lifted for the SQ, RDL, and Mean 144.0 1.72 166.9 2.00 164.1 1.96
SD 19.2 0.17 21.0 0.27 14.0 0.18
BHT was 144.0 6 19.2 (117–180) kg, 166.9 6 21.0 (145–210)

2597

Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Back Squat, Romanian Deadlift and Barbell Hip Thrust (2019) 33:10

Figure 1. Mean root mean square (RMS) values of gluteus maximus EMG activity (mean 6
SD) during maximal voluntary contraction (MVIC) and the concentric phase of the back squat
(SQ), Romanian deadlift (RDL), and barbell hip thrust (BHT) at 60 kg and one repetition
maximum (1RM) load. # Significantly different (p , 0.05) when compared with MVIC. *Sig-
nificantly different (p , 0.05) when compared with SQ. EMG 5 electromyography.

Vastus Lateralis Electromyography Discussion

For VL (Figure 3), the RMS of the SQ was greater than that of
RDL and BHT for the 60 kg (ES 5 7.77, p 5 0.002 and ES 5
2.54, p 5 0.001, respectively) as well as the 1RM load (ES 5
1.36, p 5 0.002 and ES 5 2.27, p 5 0.009, respectively). In
addition, the RMS of the RDL was smaller than that of the
BHT with the 1RM load (ES 5 0.79, p 5 0.045). In compar-
ison with the RMS during MVIC, the RDL and BHT showed
smaller RMS for the 60 kg load (ES 5 1.47, p 5 0.002 and ES
5 1.23, p 5 0.002, respectively) as well as the 1RM load (ES 5
1.11, p 5 0.003 and ES 5 0.59, p 5 0.046, respectively), but
the SQ showed greater RMS for the 1RM load (ES 5 0.90, p 5
0.006).
The results from the current study demonstrate that the BHT
elicited significantly greater GM muscle activity when compared
with the SQ but was not significantly different from the RDL at
1RM (Figure 1). There were no differences in BF activity between
the 3 exercises at 1RM (Figure 2), displaying higher activity
during all exercises with maximal loads. Vastus lateralis activity
(Figure 3) was the highest in response to the 60 kg and 1RM loads
during the SQ when compared with the RDL and BHT at 1RM
and 60 kg loads. Overall, the RMS during the SQ 1RM was
comparable with the MVIC in all muscles recorded, but the VL
activity during the RDL and BHT was lower when compared with
an MVIC (Figures 1–3). In addition, the BHT elicited higher GM

Figure 2. Mean root mean square (RMS) values of biceps femoris EMG activity (mean 6 SD)
during maximal voluntary contraction (MVIC) and the concentric phase of the back squat
(SQ), Romanian deadlift (RDL), and barbell hip thrust (BHT) at 60 kg and one repetition
maximum (1RM) load. #Significantly different (p , 0.05) when compared with MVIC. EMG 5
electromyography.

2598

Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Back Squat, Romanian Deadlift and Barbell Hip Thrust (2019) 33:10 | www.nsca.com

Figure 3. Mean root mean square (RMS) values of vastus lateralis EMG activity (mean 6
SD) during maximal voluntary contraction (MVIC) and the concentric phase of the back
squat (SQ), Romanian deadlift (RDL), and barbell hip thrust (BHT) at 60 kg and one
repetition maximum (1RM) load. #Significantly different (p , 0.05) when compared with
MVIC. *Significantly different (p , 0.05) when compared with SQ. EMG 5
electromyography.

activity in response to the 60 kg load when compared with the
other exercises. From these results, we conclude that the greater
concurrent activation in the hip and knee extensors makes the SQ
a more appropriate training exercise for sporting applications,
while the BHT and RDL are equally appropriate for isolating the
hip extensors.
The main purpose of this study was to investigate which ex-
ercise would achieve the greatest muscle activity across the 3
different muscle groups (i.e., GM, BF, and VL). Although GM
activity was greatest during the BHT, there was still high GM
activity relative to the MVIC in the 1RM for the SQ (Figure 1). In
addition, there were no significant differences in the GM activity
between 1RM and MVIC, although the loads lifted were signifi-
cantly lower in the SQ at 1RM when compared with the other
exercises. The subjects reported that the BHT was an un-
comfortable exercise to perform with maximal loads. Therefore,
although the BHT elicits the greatest GM activity, the SQ may be
a good alternative if the BHT is an uncomfortable exercise for
athletes. If the athlete has difficulty getting into the bottom po-
sition of the squat, GM activation may not be as high (4); thus, the
inclusion of BHT and RDL, along with enhancing SQ technique,
seems appropriate to ensure even strength development
throughout the anterior and posterior chain in the lower body.
Furthermore, the RDL may be an excellent alternative to the BHT
for eliciting high GM activity without the hip discomfort.
The long head of the BF works in synergy with the GM and is
involved in the end range of hip extension as well as during knee
flexion (15); thus, high hamstring activity was expected during all
3 exercises. There is no significant difference in the BF activity
between the SQ, RDL, and BHT at both the 1RM and 60 kg
loads, although higher BF RMS was seen during the BHT 1RM
when compared with an MVIC (Figure 2). The placement of the
load during the BHT across the hips supports the reliance on the
BF for full hip extension with the load applying downward force
against the hips. The BHT is performed in a supine position with
the load resisting along the sagittal plane across the hips, thus the
hamstrings were expected to elicit high activity throughout this
movement. These findings agree with Contreras et al. (8) who
reported higher BF activity during BHT when compared with the
RDL and SQ during submaximal loads. However, there were no
differences in the BF activity in the MVIC, 1RM or 60 kg loads in
all exercises, indicating high BF activity during both maximal and
submaximal dynamic contractions for the SQ, RDL, and BHT.
As a result of the 1RM loads lifted being significantly lower
during the SQ when compared with the RDL and BHT, the fixed
load of 60 kg represented a higher relative load during the SQ
than the BHT and RDL. This difference in the relative intensity
was not reflected in RMS values, as the BHT displayed higher
average GM RMS values than the RDL and SQ at 60 kg, despite
60 kg load representing a lower relative intensity during the BHT.
The ability to achieve higher activation of the GM with lower
intensities may be a useful property of the BHT. To increase the
joint compressive forces during knee flexion/extension exercises,
antagonistic activities of flexors have been previously reported
(1), which could explain the lack of difference in BF activity be-
tween the SQ and MVIC. Although the BHT was the only exercise
to elicit higher BF activation than an MVIC, there were no sta-
tistically significant differences between exercises.
During BHT 1RM testing, the subjects often reported in-
ability to go higher in weight not because of lack of strength,
but due to the discomfort caused by the weight on the hip
crease. Because of the load placement on the hips during the
BHT, this exercise poses to be an uncomfortable movement to
perform with maximal loads, even with a barbell pad. There-
fore, the high activation of the GM achieved during the BHT
without the use of heavy loads may be more conducive to hy-
pertrophic or endurance training. The lack of difference seen in
GM activity when compared with the MVIC and the 60 kg
BHT (Figure 1) may be a result of the supine position that the
exercise was performed in, which allows for an isometric
contraction of the GM at the top of the movement against the
load. This position, where there is full hip extension, is com-
parable with the position used in the MVIC measure, in which
the subjects contracted the GM against the dynamometer’s

2599

Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
 Back Squat, Romanian Deadlift and Barbell Hip Thrust (2019) 33:10
attachment arm at 180° of hip extension and may provide the
same level of activity as the top of the BHT.
In addition to the GM, the knee extensors also play an im-
portant role in functional movements of the lower body such as
running and jumping. As shown in Figure 3, VL RMS was sig-
nificantly higher with a “large” effect size during the SQ when
compared with the BHT and RDL. Vastus lateralis activity during
an MVIC was significantly higher than the activity observed in
a 1RM RDL and BHT, indicating little movement through the
knee joint during these 2 exercises. The RDL and BHT both focus
on lifting the load using the hip, with the knee in a fixed position
and the knee extensors providing stability throughout the
movement. In addition, the movement during the BHT starts in
a position of approximately 90° of knee flexion, and this angle is
maintained throughout the exercise. The SQ, however, uses the
hip as well as the knee extensors to lift the load from a position of
full flexion to full extension, demonstrated by the SQ showing
higher VL activity during 1RM when compared with MVIC.
Vastus lateralis activity was also significantly higher in the SQ
when compared with the BHT and RDL at 1RM and 60 kg.
Therefore, as the SQ displays high activity of the knee and hip
extensors, it may be a more appropriate exercise to use when
training to improve athletic movements involving simultaneous
extension of both the knee and hip.
This study showed significant differences in the BHT and the
SQ for VL activity between dynamic and isometric contractions,
which disagrees with previous findings by Contreras et al. (8). In
the study by Contreras et al. (8), “manual resistance” was pro-
vided, which may not represent adequate or consistent fixation to
evoke a true maximal isometric contraction of the VL and may
underestimate muscle activity (14). By contrast, the current study
showed no difference between RMS collected during an MVIC
and the SQ performed with 60 kg and higher activation during
submaximal dynamic contractions relative to an MVIC. This
suggests that the SQ was the only exercise to elicit high levels of
VL activation when compared with the other exercises at 1RM. In
addition, while the BHT and RDL are good for isolating the hip
extensors (i.e., BF and GM), high activity from the VL was not
seen in both exercises. Furthermore, a comparison of muscle ac-
tivity with a true MVIC measure is likely more appropriate for
practitioners looking to determine muscle activation, based on the
flaws and limitations previously highlighted.
Although the SENIAM guidelines were followed, higher levels of
activity during an MVIC when compared with submaximal con-
tractions were expected, but this was not always the case. The high
activity during submaximal 60 kg dynamic contractions, seen in the
GM, VL, and BF (Figures 1–3) when compared with an MVIC,
highlights potential differences in activation between isometric and
concentric contractions and may warrant the need for a different
normalization method in further research. In addition, while the knee
extension angle in this study was appropriate for maximal activation
(14), the same angle was used to measure hamstring activation
during knee flexion, which may have underestimated hamstring
activity. It is previously stated by Konrad (14) that systematic re-
search studies on the effectiveness of MVIC positions for maximal
muscular activation are still missing; thus, the positions used for the
normalization protocol may be the reason for resistance training
activities exceeding the established threshold for 100% activation.
In conclusion, the BHT elicited the highest GM activity with
the 1RM load when compared with the SQ, but all 3 exercises
were able to elicit high levels of GM activity during 1RM loads
relative to an MVIC. Notably, the BHT achieved the highest GM
activation even without the need for a near-maximal external
2600
Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.
load. The BHT eliciting high GM activation agrees with previous
findings by Andersen et al. (2) and Contreras et al. (8). However,
although the BHT and RDL elicit higher hip extensor activity, the
SQ was the only exercise to simultaneously provide high knee and
hip extensor activity relative to an MVIC. The SQ simultaneously
elicited high activation of the hip and knee extensors at higher
loads, which may translate better to athletic performance and
functional movement patterns that require forceful hip and knee
extension, a conclusion supported by Bishop et al. (3), previously
displaying no improvements in sprinting performance after using
the BHT for 8 weeks. This study demonstrated that while the
BHT elicited greater GM activity when compared with the SQ, no
significant differences were seen between the BHT and RDL at
1RM. Therefore, the SQ seems to be an efficient exercise for
training both hip and knee extension strength.
Practical Applications
This study demonstrates that while the BHT elicits greater
GM activity when compared with the SQ, equal levels of ac-
tivity were seen between the BHT and RDL at 1RM and the
SQ still displayed high GM activity relative to an MVIC.
Strength and conditioning coaches who aim to isolate hip
extensor musculature activation may equally benefit from
prescribing the BHT and RDL, although the BHT is usually
perceived as the safer and easier to learn option (16). How-
ever, the SQ displayed the highest levels of knee extensor ac-
tivity as well as high hip extensor activity relative to an MVIC
at 1RM loads, which places the SQ as a fundamental exercise
to be part of any resistance training program involving
movements that encompass both knee and hip extension. In
addition, the ability for the BHT to elicit higher neuromus-
cular activation of the GM with a submaximal load of 60 kg is
an important finding, as the BHT poses to be an uncomfort-
able movement to perform with maximal loads. If the goal is
to elicit high GM activity, heavy loads during the BHT may
not be necessary to accomplish it.
References
1. Aagaard P, Simonsen E, Andersen J, et al. Antagonist muscle coactivation
during isokinetic knee extension. Scand J Med Sci Sports 10: 58–67, 2000.
2. Andersen V, Fimland MS, Mo DA, et al. Electromyographic comparison
of barbell deadlift, hex bar deadlift, and hip thrust exercises: A cross-over
study. J Strength Cond Res 32: 587–593, 2018.
3. Bishop C, Cassone N, Jarvis P, et al. Heavy barbell hip thrusts do not effect
sprint performance: An 8-week randomized-controlled study. J Strength
Cond Res, 2017. Epub ahead of print.
4. Caterisano A, Moss RF, Pellinger TK, et al. The effect of back squat depth
on the EMG activity of 4 superficial hip and thigh muscles. J Strength
Cond Res 16: 428–432, 2002.
5. Clark DR, Lambert MI, Hunter AM. Muscle activation in the loaded free
barbell squat: A brief review. J Strength Cond Res 26: 1169–1178, 2012.
6. Cohen J. Statistical Power Analysis for the Behavioral Sciences (2nd ed.).
Hillsdale, MI: Erlbaum Associates, 1988.
7. Contreras B, Cronin J, Schoenfeld B. Barbell hip thrust. J Strength Cond
Res 33: 58–61, 2011.
8. Contreras B, Vigotsky AD, Schoenfeld BJ, Beardsley C, Cronin J. A
comparison of gluteus maximus, biceps femoris, and vastus lateralis EMG
activity in the back squat and barbell hip thrust exercises. J Appl Biomech
31: 452–458, 2016.
9. Contreras B, Vigotsky AD, Schoenfeld BJ, et al. Effects of a six-week hip
thrust vs. front squat resistance training program on performance in ad-
olescent males: A randomized controlled trial. J Strength Cond Res 31:
999–1008, 2017.
 Back Squat, Romanian Deadlift and Barbell Hip Thrust (2019) 33:10 | www.nsca.com

10. Earle RW. Weight training exercise prescription. In: Essentials of Personal
Training Symposium Workbook. Lincoln, NE: NSCA Certification
Commision, 2006.
11. Faul F, Erdfelder E, Lang AG, Buchner A. G* power 3: A flexible statistical
power analysis program for the social, behavioral, and biomedical sci-
ences. Behav Res Methods 39: 175–191, 2007.
12. Frounfelter G. Teaching the Romanian deadlift. J Strength Cond Res 22:
55–57, 2000.
13. Hermens H, Freriks B, Merletti R, et al. European Recommendations for
Surface Electromyography. Enschede, the Netherlands: Roessingh Re-
search and Development, 1999. pp. 13–54.
14. Konrad P. The ABC of EMG: A Practical Introduction to Kinesiological
Electromyography. Scottsdale, AZ: Noraxon, 2005. pp. 30–35.
15. Kwon YJ, Lee HO. How different knee flexion angles influence the
hip extensor in the prone position. J Phys Ther Sci 25: 1295–1297,
2013.
16. Lane C, Mayer J. Posterior chain exercises for prevention and treatment of
low back pain. ACSMs Health Fit J 21: 46–48, 2017.
17. Luca CJD. The use of surface electromyography in biomechanics. J Appl
Biomech 13: 135–163, 1997.
18. Neumann DA. Kinesiology of the hip: A focus on muscular actions.
J Orthop Sports Phys Ther 40: 82–94, 2010.
19. Parker J, Connors B. Indiana university basketball inseason strength
program. J Strength Cond Res 1: 5–6, 1979.
20. Whaley O, McClure R. Another perspective on teaching the pulling
movements. J Strength Cond Res 19: 58–61, 1997.

2601

Copyright © 2019 National Strength and Conditioning Association. Unauthorized reproduction of this article is prohibited.